No-Web: The Inevitable Future of Digital Content?
Exploring the AI Impact on the Digital Publishing Business
The digital revolution has shaken up the media world like never before over the past few decades. It significantly changed how we consume news, music, and video. The shift began with the dematerialization of print and other tangible media then continued with the unbundling of articles from newspapers, songs from albums, and video from cable networks. This progression has been a rollercoaster ride for media companies, forcing them to reinvent their business model in a fully digital world. Yet, just when they thought they had it figured out, a new wave of turbulence is on the horizon, one that will be more profound and brutal than the first.
The rise of ChatGPT has been phenomenal, to say the least. The staggering estimate of 100 million users, just two months after its launch, testifies to the unparalleled level of interest. This kind of scale leaves little room for doubt:
People are strongly drawn to conversational interactions for many types of queries.
A few exchanges are all it takes to realize - there's simply no going back. The direct responses to questions, presented in a pristine manner, feel like a breath of fresh air in a polluted web with SEO junk.
In the present world of written media, distribution has become highly dependent on standalone articles and the ability of their headlines to generate clicks. As a result, the search engine became a strong driving force behind media exploration. The value is no longer derived from the worth of an audience within a branded publication, but instead from the number of clicks each article could generate. As users increasingly turn to pre-trained AI chat tools for their queries, one may wonder: what will happen to the online publishing business?
There are essentially two types of search queries. The first type involves queries that seek direct answers or solutions to specific issues. The second category contains queries that aim to uncover source websites for more in-depth exploration and research. It may appear that the latter type is less impacted at the moment, as people tend to revert to traditional search methods to access additional content. But is this truly the case?
Despite its popularity, ChatGPT has a couple of key limitations. On the one hand, it's confined to the knowledge available before its training data cutoff date, which was at the end of 2021. On the other hand, it operates in an isolated environment from the Internet, which prevents it from accessing real-time updates and information.
That said, some companies have already begun offering solutions to overcome these limitations. Microsoft Bing, Neeva, and Perplexity.ai are a few of the players trying a combination of traditional search with AI-powered chat.
The current hybrid approaches may sound appealing at first. Still, they are just transition solutions as companies are trying to experiment with AI-powered search without destroying the existing business models. In other words, these crossbreed offerings are attempting to safeguard "the link" at the cost of a subpar user experience. Do you really want your search results to be a mishmash of both answers to your query and a wall of blue links? It's like a lifejacket for the status quo, sacrificing the smooth sailing of the user experience for a false sense of security.
Now, compare the screenshot above with this one:
As a starting point, which experience would you rather have?
What we see Microsoft and other players now launching is the typical initial attempts of trying to fit "the old" inside "the new." Looking back at the early iPad days, Apple went through the same process. For example, when they released Newsstand, they tried to put fancy PDF versions of paper magazines inside an App.
The skeuomorphic design imitating the appearance of a real-world stand looked cool, but the approach of creating isolated digital replicas of print magazines ultimately flopped.
Apple eventually realized they had to rebuild the concept from scratch and tailor it for the new digital reality. So, four years later, they replaced Newsstand with Apple News. Through a single subscription, users could access their favorite magazines with a far better experience that feels more native to the digital format. The App also offered a personalized "For You" feed featuring a blend of articles from various publications.
As we examine how the mixture of search and AI chat is done in the present solutions, it appears we may be having a "Newsstand moment."
I believe, there are far superior ways to implement a combination of traditional search with AI chat. What we have seen so far feels like an unappetizing cocktail with ingredients not blending together.
The No-Web Future
Imagine a future where chat-style interfaces are at the forefront. No more scrolling through endless results bloated with ads. Instead, these interfaces would abstract away the external search by distilling the information from results down to its core, then presenting it in an easy-to-digest form inside an uncluttered chat interface.
Here's how it could work:
Based on the user's query, the system will determine whether to utilize its AI language model or a traditional search engine. If the pre-trained AI model can answer the question, it'll operate similarly to the current ChatGPT. But if the inquiry requires an external search, such as "What's the latest news from the elections?"the system will tap into a standard search engine API. However, rather than polluting the user experience with a wall of blue links, it'll first send them to the AI language model for further analysis and summarization before presenting the distilled information to the end user. This approach would offer the best of both worlds, providing a refined experience without sacrificing quality and accuracy.
Now, here's the dilemma. Even though such a service will be best for consumers, it will also have a devastating impact on the publishing industry. It will dismantle "the link," ultimately causing a drop in traffic to the source publications, leading to a decrease in display ads and a drying up of the financial lifeline for content creators.
Looking at the decline in newspaper ad revenue over the past two decades, we see that even as the industry has shifted to digital ads, it remains a far cry from its peak in 2000. The rise of Google, and later Facebook, has dramatically altered the ad market by creating intent-based and personalized targeting, allowing brands to reach their audiences with greater precision.
It's an ironic twist of fate that the very tech giants that stole the advertising business from traditional publications have also become their saving grace by driving traffic and keeping them afloat. Imagine the magnitude of the revenue drop if newspapers didn't have the traffic from Google. It appears the moment has arrived where we'll witness this scenario starting to play out.
Even though traditional pages may still be around in the future, they will significantly decrease in relevance. Chat-based interfaces will curate and present information in a personalized, conversational format. Instead of browsing pages, users will spend more time in Q&A mode, and we'll progressively move to a No-Web world.
Monetizing Content in a No-Web World
In that new reality, publishers will have to solve the puzzle of how to remain relevant, keep their content in front of their audience, and generate revenue. The question on everyone's mind is:
Should they ringfence their content behind a paywall or put their faith in the promise of citation links?
The harsh truth is that a mere mention at the end of a chat session is unlikely to drive a level of traffic anywhere close to a top Google search result. If users can already get what they need, why bother clicking on that link?
At this stage, it is hard to predict exactly how the ecosystem will evolve but we can explore a couple of potential developments that might occur in combination.
People consume content in two distinct ways: "pull mode" and "push mode." In "pull mode," individuals actively hunt for information using keywords and search engines. This is the mode that AI-powered front-end chat interfaces will ultimately disrupt. In "push mode," individuals passively receive information, relying on the suggestions of curation services such as Apple News, Flipboard, and, more recently, Artifact. These personalized feeds offer a tailored experience, and publishers have the opportunity to double down on this mode by removing their content from the open web and making it available only through a subscription.
It is essential though that a large number of publications are accessible through a single subscription for the end user. That's the model that has saved the Music industry.
As mp3s dematerialized CDs, sales collapsed. A glimmer of hope emerged with legal download services such as iTunes. Still, it wasn't until music streaming offered the entire catalog with a single subscription that the industry could reclaim its former glory. Music has now jumped back to its peak level of revenue1, last seen in the late 1990s.
The alliances with the news aggregation services outlined earlier play a vital role in the "push mode." Nevertheless, for publishers to truly succeed in the new No-Web landscape, they must also figure out a strategy to prevent the "pull mode" from being eaten by AI-powered chatbots.
Consider this perspective: What if AI-powered clients were actually a golden opportunity waiting to be leveraged to great effect rather than a looming danger to be feared?
Instead of standing still and watching their traffic and ad revenue dry up, publishers have a unique opportunity to forge lucrative B2B deals with tech giants like Microsoft and Google. By removing their pages from the public web and granting special access to their prized content to these AI chat clients, publishers can tap into a new revenue stream without having to sell to millions of individual consumers. This could be a far more effective monetization strategy, as they would only need to secure a limited number of large deals with select tech companies. The responsibility for charging end users with a bundled subscription will then fall upon the search and AI providers.
Today, those kinds of arrangements are not feasible. Search engines have no incentives to pay. The ongoing deal is "free traffic" in exchange for "free content." But if "free traffic" were to disappear, something else would need to be offered as compensation for content.
Final Thoughts
The dismantling of the open web publishing business model, as we know it, is most likely inevitable. What remains unclear is the speed at which it will happen and what the final outcome will look like.
In this article, the focus was explicitly on written digital content. It is essential to highlight, though, that the shift we're seeing towards AI chatbots will not be just limited to the realm of publishing. It's a change that's set to ripple across the entire web, from booking and retail to ticketing. Websites currently serving as these services' hubs will progressively migrate towards API-based platforms that AI chat clients will tap into. Yet, the distinction between this scenario and content publishing is that a drop in traffic to these services' traditional websites won't at all affect their business models and revenue streams. For instance, if a ticketing platform generates revenue through commissions, it doesn't matter if the order originates from a website or an AI assistant. To these services, AI chatbots will only serve as another channel, providing consumers with more options and versatility.
If we try to guess how things may end up at a broader level, it's possible that the Internet will undergo a radical transformation. It might emerge as a more invisible version of its current form. The bulk of websites and apps that we use today may be deconstructed down to the two components below:
The seemingly innocuous Spotlight Search and Siri are often ridiculed as they cannot do much. Spotlight is primarily good at in-device search, and Siri is mostly good at setting timers. But what if these unremarkable tools were armed with the power of a GPT-like AI backend, capable of distilling results from full-fledged search and accessing a multitude of API-based services? The ramifications would be immense, as billions of iOS devices around the globe would suddenly gain the ability to open the tap on the new No-Web reality and, at the same time, switch off the Google-dominated world as we know it2.
Even though the revenues are back to their historical peak level, the breakdown of the pie is very different than in the 90s. The value is now highly concentrated between the major record labels and the streaming platforms. The cut of artists has dropped significantly. But that's a discussion for another post.
The example was given based on Apple iOS Spotlight and Siri. But other platforms like Microsoft Bing and new independent players will most likely evolve in a similar direction.
Additional Reading
When The Internet Becomes Chat | Troy Young
Why don’t Search Engines Integrate ChatGPT-like Bots in More Helpful Ways | Gina Chua
Podcast
The New Media Interface | Troy, Alex and Brian on People vs Algorithms